Merge pull request #412 from deepgram/feat/diarize-model-v2

feat: add diarize_model option to pre-recorded transcription

Author: GregHolmes

.agents/skills/deepgram-dotnet-audio-intelligence/SKILL.md

@@ -96,7 +96,7 @@ await liveClient.Connect(new LiveSchema()
 
 ## Key params
 
-REST (`PreRecordedSchema`): `Summarize`, `Topics`, `Intents`, `Sentiment`, `DetectLanguage`, `DetectEntities`, `CustomTopic`, `CustomTopicMode`, `CustomIntent`, `CustomIntentMode`, `Diarize`, `DiarizeVersion`, `Redact`, `Utterances`, plus the regular STT knobs.
+REST (`PreRecordedSchema`): `Summarize`, `Topics`, `Intents`, `Sentiment`, `DetectLanguage`, `DetectEntities`, `CustomTopic`, `CustomTopicMode`, `CustomIntent`, `CustomIntentMode`, `Diarize`, `DiarizeVersion`, `DiarizeModel`, `Redact`, `Utterances`, plus the regular STT knobs.
 
 Live (`LiveSchema`): `Diarize`, `Redact`, `UtteranceEnd`, `VadEvents`, `Punctuate`, `SmartFormat`, plus normal streaming STT settings.
 

Deepgram.Tests/UnitTests/UtilitiesTests/QueryParameterUtilTests.cs

@@ -182,4 +182,36 @@ public void GetParameters_Should_Return_Empty_String_When_Parameter_Has_No_Value
         result.Should().Be($"https://{Defaults.DEFAULT_URI}");
     }
 
+    // DiarizeModel surfaces Batch Diarization v2; the API spec enum is latest, v1, v2.
+    [TestCase("v2")]
+    [TestCase("v1")]
+    [TestCase("latest")]
+    public void GetParameters_Should_Return_String_When_Passing_DiarizeModel_Parameter(string diarizeModel)
+    {
+        // Input and Output
+        var obj = new Deepgram.Models.Listen.v1.REST.PreRecordedSchema() { DiarizeModel = diarizeModel };
+        var expected = $"diarize_model={diarizeModel}";
+
+        //Act
+        var result = QueryParameterUtil.FormatURL(Defaults.DEFAULT_URI, obj);
+
+        //Assert
+        result.Should().NotBeNull();
+        result.Should().Contain(expected);
+    }
+
+    [Test]
+    public void GetParameters_Should_Omit_DiarizeModel_When_Not_Set()
+    {
+        // Input and Output - additive/non-breaking: absent unless explicitly set
+        var obj = new Deepgram.Models.Listen.v1.REST.PreRecordedSchema() { Model = "nova-2" };
+
+        //Act
+        var result = QueryParameterUtil.FormatURL(Defaults.DEFAULT_URI, obj);
+
+        //Assert
+        result.Should().NotBeNull();
+        result.Should().NotContain("diarize_model");
+    }
+
 }

Deepgram/Models/Listen/v1/REST/PreRecordedSchema.cs

@@ -91,6 +91,17 @@ public class PreRecordedSchema
 	[JsonPropertyName("diarize")]
     public bool? Diarize { get; set; }
 
+    /// <summary>
+    /// Select and enable a specific batch diarization model version, e.g. "latest", "v1", or "v2".
+    /// Supersedes the deprecated <see cref="Diarize"/> boolean; if set, do not also set Diarize.
+    /// Batch/pre-recorded only - not accepted on streaming requests.
+    /// <see href="https://developers.deepgram.com/docs/diarization">
+    /// default is null
+    /// </summary>
+    [JsonIgnore(Condition = JsonIgnoreCondition.WhenWritingNull)]
+	[JsonPropertyName("diarize_model")]
+    public string? DiarizeModel { get; set; }
+
     // <summary>
     /// <see href="https://developers.deepgram.com/docs/diarization">
     /// default is null, only applies if Diarize is set to true

Deepgram/README.md

@@ -293,6 +293,7 @@ also see [TranscribeSchema] which LiveSchema is derived from for more options
 | Callback               | string?                   | Callback URL to provide if you would like your submitted audio to be processed asynchronously                          |
 | Diarize                | bool?                     | Indicates whether to recognize speaker changes                                                                         |
 | DiarizeVersion         | string?                   |                                                                                                                        |
+| DiarizeModel           | string?                   | Batch diarization model version to use (latest, v1, v2). Supersedes the deprecated Diarize boolean                     |
 | Extra                  | Dictonary<string,string>? |  additonal values you want echoing bac                                                                                 |                                                                                                                        |
 | FillerWords            | int?                      | Whether to include words like "uh" and "um" in transcription output.                                                   |
 | Keywords               | List<string>?             | Keywords to which the model should pay particular attention to boosting or suppressing to help it understand context   |